Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-10-04
13:43:19 openstackgerrit Surya Seetharaman proposed openstack/nova master: [WIP] Refactor scatter-gather utility to return exception objects https://review.openstack.org/607934
13:50:13 openstackgerrit Jack Ding proposed openstack/nova master: Handle missing marker during online data migration https://review.openstack.org/605164
14:02:19 dansmith did I miss that the meeting this week was canceled or something?
14:03:06 gibi dansmith: I think I did not get a request from melwitt to chair it but I can
14:03:25 dansmith meh
14:03:26 gmann not sure. I was also checking meeting channel
14:03:32 efried we have an important topic to discuss, HPET for libvirt.
14:03:34 bauzas hah, I just asked the question in -meeting
14:03:36 efried they're expecting it.
14:03:45 bauzas efried: and what about HCATTLE ?
14:03:56 efried oy vay
14:04:06 sean-k-mooney efried: so #openstack-meeting time i guess
14:04:24 efried yup
14:11:10 mrch_ can i prevent the nova.compute.resource_tracker spaming this: nova.compute.resource_tracker [req-e23cf459-506f-42d3-ac5d-73d87b1e5ba9 - - - - -] [instance: 8ecbbdb1-bad7-434b-8d09-265964ca9c7e] Instance not resizing, skipping migration. when 5 hosts spam this about one an the same vm, an i got 10 vms of that kind, its annoying as f
14:12:57 efried mrch_: We're in the nova meeting atm; can you ask again at the top of the hour?
14:13:43 efried It looks like this is happening because you have some stale migrations in the database; can you clean those up?
14:15:56 mrch_ efried
14:15:56 mrch_ efried: shure
14:18:05 mrch_ efried
14:18:06 mrch_ efried: concidering db, but i cant wait, i stay in chat just write ill see it tomorrow
14:18:50 efried mrch_: Okay. I personally don't know how to clean up the db, but hopefully one of the wizards can advise once the meeting is over.
14:58:22 melwitt gibi: thank you for running the nova meeting, it completely slipped my mind this time. apologies
15:00:26 gibi melwitt: no worries
15:00:35 mhen o/
15:01:46 mrch_ can i prevent the nova.compute.resource_tracker spaming this: nova.compute.resource_tracker [req-e23cf459-506f-42d3-ac5d-73d87b1e5ba9 - - - - -] [instance: 8ecbbdb1-bad7-434b-8d09-265964ca9c7e] Instance not resizing, skipping migration. when 5 hosts spam this about one an the same vm, an i got 10 vms of that kind, its annoying as f
15:02:37 mhen as already mentioned in the ML we want to propose the introduction of image encryption in OpenStack
15:02:42 mhen #link ML thread on proposal for image encryption http://lists.openstack.org/pipermail/openstack-dev/2018-September/135167.html
15:02:53 mhen as advised on the ML, we are currently writing individual specs for the components involved, including Nova
15:03:01 mhen in Nova we would like to add the decryption of images, which are converted into servers using ephemeral storage and the encryption of images created from such servers
15:03:22 mhen at first we want to focus on LibVirt and its LVM backend, because it is the only backend for encrypted ephemeral storage in LibVirt (ignoring the ineffectiveness of the encryption due to missing native LUKS support for now)
15:03:45 mhen we would be interested in Nova's opinion on whether the conversion from encrypted image to unencrypted ephemeral storage should be allowed
15:03:58 mhen we don't really like the idea of exposing the image data this way but the Cinder team suggested making this optional for the similarly natured volume use case
15:04:36 efried jaypipes: Because then the libvirt driver would have to use that (HPET=required) as its cue to turn HPET on in the guest. I.e. using a trait to effect configuration.
15:05:23 efried jaypipes: I'm writing up the alternatives with their respective cons; stay tuned.
15:05:56 cfriesen sean-k-mooney: if we ever wanted to support passthrough of host TPM, then there really would be a finite inventory
15:06:26 cfriesen sean-k-mooney: for emulated TPM we could make it a trait if the host supports it
15:08:44 cfriesen sean-k-mooney: if we ever wanted to support both emulated and passthrough TPM, what would the placement side look like? would that be two resource providers, one with the "emulated" trait and one with the "real" trait?
15:09:13 jaypipes mrch_: why would 5 hosts be spamming that same message? does that mean you have the same instance stuck in a migration to multiple hosts?
15:09:44 jaypipes cfriesen: and that's the problem with embedding key/value things into a trait string...
15:10:50 cfriesen jaypipes: maybe no traits are needed....in a sense this would be equivalent to VCPU vs PCPU
15:11:07 cfriesen so maybe it's just two separate resource providers
15:13:46 jaypipes cfriesen: ack
15:23:33 spatel Quick question, i am creating SR-IOV instance and in order to do that i need to create neutron port first. i have created bunch of port during testing but now i need to find out which neutron port is active and non-active how do i find that out?
15:24:07 spatel sean-k-mooney: morning
15:24:16 openstackgerrit Eric Fried proposed openstack/nova-specs master: WIP: High Precision Event Timer (HPET) on x86 guests https://review.openstack.org/607989
15:24:48 efried sean-k-mooney, cdent, jaypipes, cfriesen, dansmith, bauzas, gibi: ^
15:24:56 cdent noted
15:25:41 dansmith jaypipes: you around for a hangout in 35 minutes on this hpet thing?
15:32:50 cfriesen efried: got it. thanks for writing this up
15:36:17 jaypipes dansmith: unfortunately, no. I would be available in about 1.75 hours though
15:36:44 dansmith efried: cfriesen cdent bauzas gibi ^ ?
15:37:00 bauzas sorry I'm trapped in a meeting, trying to catch up
15:37:08 bauzas ah
15:37:16 bauzas I'm fine with meeting us later
15:37:25 efried I can't make it then :(
15:37:27 bauzas the best for me is after 8pm UTC
15:37:36 cdent dansmith: I will have to pass, but I trust the combo of people involved to work out something sane, especially since you were able to identify what is sacred to me (and remains so)(
15:37:52 dansmith heh okay
15:38:39 efried 8pm UTC would wfm
15:38:48 jaypipes that works for me too
15:38:57 cfriesen so a little over 4hrs from now?
15:39:25 efried yes
15:39:25 cfriesen works for me
15:40:35 cfriesen I've never done one of the hangouts....does someone post a URL or send out invites or what?
15:40:51 dansmith url
15:41:01 jackding I will have to pass.
15:41:13 efried cfriesen: dansmith will drop a URL in here, ping all of us, and get spanked by Sigyn
15:41:15 cfriesen I'll cover for jack
15:58:30 openstackgerrit Surya Seetharaman proposed openstack/nova master: Return a minimal construct for nova show when a cell is down https://review.openstack.org/591658
16:02:29 efried gibi: You still around?
16:02:44 gibi efried: yeah for about 10 minutes
16:02:49 gibi efried: but go ahead
16:03:01 efried okay, that may not be enough. I wanted to walk through the allocation_request_version nonsense a bit.
16:03:41 gibi efried: even if we figure out that we want to drop allocation_request_version I would like to do that separately as that a surgery again
16:04:06 efried it's just, I see it as being pretty much unnecessary/unrelated to this patch.
16:04:12 openstackgerrit Balazs Gibizer proposed openstack/nova master: Ignore forcing of live migration for nested instance https://review.openstack.org/605785
16:04:13 openstackgerrit Balazs Gibizer proposed openstack/nova master: Consider nested allocations during allocation cleanup https://review.openstack.org/606050
16:04:13 openstackgerrit Balazs Gibizer proposed openstack/nova master: Ignore forcing of evacuation for nested instance https://review.openstack.org/606111
16:04:14 openstackgerrit Balazs Gibizer proposed openstack/nova master: Run negative server moving tests with nested RPs https://review.openstack.org/604125
16:04:30 efried I guess that alone isn't worth holding it up since jay is +2 at this point.
16:04:40 efried And I guess I do see that it becomes necessary again for nested.
16:04:48 efried it's just a weird transition.
16:06:18 gibi efried: totally agree that allocation_request_version is strange but I'm not 100% sure we simply drop it and fix the version in the report client globally
16:06:57 efried gibi: IIUC, it's necessary whenever there's a possibility of the conductor and the scheduler being at different versions.
16:07:40 efried ...versions of nova, leading to them being aware of different versions of placement. Placement ought to always be at least up to date with the later of the two.
16:08:07 efried I assume that's still a possible scenario, for like rolling upgrades or whatever.
16:08:39 gibi efried: I'm not sure we support that scheduler and conductor on different version. What complicates it is that we send alternates around in RPC messages for reschedule
16:08:48 efried I'm guessing there's a rule/guideline, like you have to upgrade the computes before the scheduler, or vice versa? dansmith do you know that answer?
16:09:04 gibi efried: controller first is the ruel
16:09:05 gibi rule
16:09:35 gibi but I think controller services needs to be updated at once but I'm not sure
16:09:48 gibi I have to run now. talk to you tomorrow
16:10:07 efried thx gibi
16:10:28 gibi efried: thank you
16:13:21 openstackgerrit Merged openstack/nova master: Skip test_parallel_evacuate_with_server_group until fixed https://review.openstack.org/607620
16:13:55 openstackgerrit Jack Ding proposed openstack/nova master: Handle missing marker during online data migration https://review.openstack.org/605164
16:17:36 openstackgerrit Surya Seetharaman proposed openstack/nova master: Return a minimal construct for nova service-list when a cell is down https://review.openstack.org/584829
16:23:11 dansmith efried: control services first and together, computes last and in any order you want
16:23:49 efried dansmith: Okay, so the reason for allocation_request_version is in case computes are currently downlevel from controller and/or each other. Thanks.
16:24:00 openstackgerrit Jack Ding proposed openstack/nova master: Add HPET timer support for x86 guests https://review.openstack.org/605902

Earlier   Later